Methodology

Overview

LumiKin measures two things about every game and reports them independently. The first is what a child stands to gain from playing — measured as a Benefits Development Score (BDS) across cognitive, social-emotional, and motor dimensions. The second is the design pressure the game exerts on a developing brain — measured as a Risk Influence Score (RIS) across dopamine manipulation, monetization pressure, and social risk. From these two scores, LumiKin derives two outputs: a LumiScore (0–100), which is a harmonic mean of benefit and safety that penalises games that are high-risk or low-benefit on either axis; and a daily time recommendation in five tiers, derived from the RIS with adjustments for unusually high or low benefit scores. Many games score meaningfully on both axes simultaneously, and LumiKin is designed to make that visible rather than collapse it into a single judgement.

LumiKin does not measure how much a child enjoys a game, how skilled a player is, or whether a particular title is "good" or "bad" in any general sense. It measures structural design choices — what the game asks of a player and what mechanisms it uses to keep them playing. A beloved family favourite can carry meaningful design risks; a forgettable puzzle game can quietly deliver real cognitive benefit. The framework is deliberately separated from taste, popularity, and critical reception, all of which are well-served by other rating systems.

The framework is designed to be auditable. Every dimension has a fixed scoring scale, every category has a documented weight, and every score carries the methodology version that produced it. When the rubric changes, a new version is published and prior versions remain queryable. Reviewers — whether human or automated — score against the same rubric, with reference titles used to calibrate against drift. Limitations and edge cases (UGC platforms, live-service content, automated scoring) are documented openly in the Limitations section below rather than buried.

The Scoring Model

Every LumiKin game profile produces two independent composite scores and a time recommendation derived from them.

Benefits Development Score (BDS)

The BDS combines three benefit categories using fixed weights:

Each category score is normalised to 0–1 before weighting.

Formula: BDS = (B1_norm × 0.50) + (B2_norm × 0.30) + (B3_norm × 0.20)

Risk Influence Score (RIS)

The RIS combines three risk categories. Content risk (R4) is reported separately and does not feed into the time recommendation — this preserves compatibility with ESRB and PEGI ratings.

Formula: RIS = (R1_norm × 0.45) + (R2_norm × 0.30) + (R3_norm × 0.25)

Time recommendation tiers

The primary output is a recommended daily time limit derived from the RIS, with adjustments for high-benefit and low-benefit games.

Tier adjustments:

• If BDS ≥ 0.60 and RIS ≤ 0.70: recommendation extends one tier (e.g. 60 → 90 min)
• If BDS < 0.20 and RIS > 0.30: recommendation drops one tier (e.g. 60 → 30 min)

These adjustments do not stack.

B1 Cognitive

Rationale: Cognitive development is the largest single weight in the BDS (50%) because the age range LumiKin targets — roughly six to sixteen — overlaps with significant maturation of executive function, working memory, spatial reasoning, and metacognition. Games occupy a meaningful share of discretionary time in this window. The mechanics a child engages with repeatedly during this period are not neutral: they constitute a form of practice. Some games offer rich, varied cognitive challenges; others offer narrow, repetitive ones. The B1 dimensions attempt to characterise that landscape so parents and partners can distinguish between superficially similar titles that differ substantially in what they actually train.

What's measured: Ten dimensions, each scored 0–5 by a trained reviewer:

Scoring: Sum of 10 dimensions (0–5 each). Max 50. Normalised to 0–1 for weighting.

Weighting: B1 contributes 50% of the BDS — the largest single weight, reflecting the centrality of cognitive development in the age ranges LumiKin targets.

Research basis:

• Bavelier, D. & Green, C.S. (2019). Enhancing attentional control: Lessons from action video games. Neuron, 104(1), 147–163. [verify]
• Granic, I., Lobel, A., & Engels, R.C.M.E. (2014). The benefits of playing video games. American Psychologist, 69(1), 66–78. [verify]
• Sala, G., Tatlidil, K.S., & Gobet, F. (2018). Video game training does not enhance cognitive ability: A comprehensive meta-analytic investigation. Psychological Bulletin, 144(2), 111–139. [verify] (Included deliberately as a critical counterweight — LumiKin does not claim broad transfer effects, and the B1 framework distinguishes specific trained skills from general cognitive enhancement.)

B2 Social-Emotional

Rationale: Cooperative play, narrative-driven games, and well-moderated multiplayer environments can develop genuine social and emotional skills — perspective-taking, emotional regulation under pressure, ethical reasoning in low-stakes contexts. The same structural feature that makes games effective at training these skills (sustained social interaction with consequence) also makes them vulnerable to exploitation, which is why the social-emotional dimensions are scored separately from social-emotional risks (R3). A game can simultaneously develop authentic teamwork (high B2) and expose a child to identity pressure or unmoderated strangers (high R3); these are independent axes and the framework treats them that way.

What's measured: Six dimensions, each scored 0–5:

Scoring: Sum of 6 dimensions (0–5 each). Max 30. Normalised to 0–1 for weighting.

Weighting: B2 contributes 30% of the BDS.

Research basis:

• Greitemeyer, T. & Mügge, D.O. (2014). Video games do affect social outcomes: A meta-analytic review of the effects of violent and prosocial video game play. Personality and Social Psychology Bulletin, 40(5), 578–589. [verify]
• Harrington, B. & O'Connell, M. (2016). Video games as virtual teachers: Prosocial video game use by children and adolescents from different socioeconomic groups is associated with increased empathy and prosocial behaviour. Computers in Human Behavior, 63, 650–658. [verify]

B3 Motor

Rationale: Motor development is the smallest weight in the BDS (20%) because the dominant concern for screen time in this age range is sedentary behaviour rather than fine-motor deficit. The dimension is retained because some games — particularly motion-controlled, VR, and AR titles — meaningfully extend into physical activity, and because hand-eye coordination training has documented co-benefits in attention and reaction tasks. Games that contribute negligibly to motor development are scored low here without penalty to the overall BDS provided they perform well on B1 and B2.

What's measured: Four dimensions, each scored 0–5:

Scoring: Sum of 4 dimensions (0–5 each). Max 20. Normalised to 0–1 for weighting.

Weighting: B3 contributes 20% of the BDS.

Research basis:

• Polechoński, J., Dębska, M., & Dębski, P.G. (2020). Exergaming can be a health-related aerobic physical activity. BioMed Research International, 2020. [verify]
• Li, R., Polat, U., Makous, W., & Bavelier, D. (2009). Enhancing the contrast sensitivity function through action video game training. Nature Neuroscience, 12(5), 549–551. [verify]

R1 Dopamine Design

Rationale: Dopamine manipulation design is weighted highest in the RIS (45%) because the mechanisms involved — variable ratio reward schedules, near-miss feedback, loss aversion triggers, time-limited urgency — are the most thoroughly documented in behavioural psychology and the most aggressively engineered into modern free-to-play games. The same techniques are used in regulated gambling products. The developmental concern is acute: the prefrontal regulation that helps adults resist these mechanisms is still maturing through adolescence. R1 attempts to characterise the intensity and variety of these mechanisms in a given title, on the principle that a game with one mild streak feature and a game built end-to-end around variable rewards present qualitatively different risks even if both technically contain "engagement design".

What's measured: Ten risk factors, each scored 0–3:

Scoring: Sum of 10 factors (0–3 each). Max 30. Normalised to 0–1 for weighting.

Weighting: R1 contributes 45% of the RIS — the highest weight, reflecting the documented developmental harm of these mechanics in children's products.

Research basis:

• Zendle, D. & Cairns, P. (2018). Video game loot boxes are linked to problem gambling: Results of a large-scale survey. PLOS ONE, 13(11). [verify]
• King, D.L. & Delfabbro, P.H. (2019). Predatory monetization features in video games (e.g., 'loot boxes') and Internet gaming disorder. Addiction, 114(11), 1967–1969. [verify]
• UK Competition and Markets Authority (2023). Online choice architecture: How digital design can harm competition and consumers. Discussion paper. [verify]
• The R1 dimension list draws on the SHARP-G framework (Saini & Hodgins, 2024) [verify] adapted for children's contexts.

R2 Monetization

Rationale: Monetization pressure (R2, 30% of RIS) is scored separately from dopamine design because the mechanisms are distinct even when they co-occur. R1 measures psychological extraction — how strongly the game pulls a child into continuing to play. R2 measures financial extraction — how strongly the game pulls a child or parent into continuing to spend. A game can be high on one and low on the other: a free advertising-funded game may have minimal monetization pressure but heavy dopamine design; a flat-fee premium title may have low dopamine design but in-game purchase prompts targeting children. R2 is particularly load-bearing in regulatory contexts (UK Children's Code, FTC guidance on dark patterns in children's apps), and partner products typically need to filter on it independently of overall game quality.

What's measured: Eight risk factors, each scored 0–3:

Scoring: Sum of 8 factors (0–3 each). Max 24. Normalised to 0–1 for weighting.

Weighting: R2 contributes 30% of the RIS.

Research basis:

• Federal Trade Commission (2022). Bringing dark patterns to light. FTC Staff Report. [verify]
• Information Commissioner's Office, UK (2020/2025). Age Appropriate Design Code (Children's Code). [verify]
• Fitton, D. & Read, J.C. (2019). Creating a framework to support the critical consideration of dark design aspects in free-to-play apps. Proceedings of IDC '19. [verify]

R3 Social Risk

Rationale: Social risk is scored independently of the positive social dimensions in B2 because they describe different phenomena even when they share a surface — multiplayer interaction. A game with strong cooperative design and active community moderation can develop genuine teamwork (high B2) while remaining low-risk (low R3). A game with the same multiplayer surface but without moderation, with public ranking, and with unmoderated voice chat available to strangers can develop similar surface skills while exposing a child to qualitatively different risks. R3 attempts to characterise those risks distinctly: the chance of toxic competitive environments, the presence of unknown adults in voice or chat, social comparison pressure tied to spending or in-game performance, and privacy practices. The most acute concern in this category — particularly for children under 13 — is unmoderated stranger contact, which is scored as its own factor and weighted heavily within R3.

What's measured: Six risk factors, each scored 0–3:

Scoring: Sum of 6 factors (0–3 each). Max 18. Normalised to 0–1 for weighting.

Weighting: R3 contributes 25% of the RIS.

Research basis:

• Kowert, R. (2020). Dark Participation in Games. Frontiers in Psychology, 11. [verify]
• Internet Watch Foundation annual reports — relevant sections on grooming via gaming platforms. [verify]
• 5Rights Foundation (2021). Pathways: How digital design puts children at risk. [verify]

R4 Content Risk

Important: R4 is reported as a standalone content flag and does not feed into the RIS or the time recommendation. This is a deliberate design decision: content risk largely aligns with existing ESRB and PEGI ratings, and LumiKin's primary contribution is in areas those ratings don't cover (manipulation design, monetization, social risk). Including R4 in the time recommendation would create overlap with regulatory systems and reduce the distinctiveness of the LumiScore.

What's assessed: Five factors, each scored 0–3:

Display: R4 scores are shown as a content advisory alongside the ESRB/PEGI rating on game pages, not as a component of the overall LumiScore.

Update and Versioning Policy

Games are rescored under three triggers. First, automated rescore is initiated when the upstream metadata source (RAWG) reports a substantive update — DLC, platform expansion, or content patches that materially alter mechanics. Second, manual rescore is initiated for live-service titles where mechanics have changed in ways the automated pipeline did not detect. Third, ad-hoc rescore is initiated when LumiKin receives credible information about changes via community reports, regulatory action, or publisher disclosure.

The scoring weights, dimension definitions, scoring scales, and tier thresholds are locked within a methodology version. Any change to these — adding a dimension, adjusting a weight, redefining a tier boundary — requires publishing a new methodology version. Version numbers follow semantic versioning: minor version increments (1.0 → 1.1) for additive changes that do not alter existing scores, major version increments (1.0 → 2.0) for changes that require all games to be rescored. Every score carries the methodology version that produced it. Historical scores remain queryable under their original version via the ?version=X parameter on this page and via the methodology_version field in the API. Partners consuming the API receive methodology_version and calculated_at fields on every score response.

Limitations and Edge Cases

UGC platforms (Roblox, Fortnite Creative). LumiKin scores individual experiences on UGC platforms as well as the platform itself. The platform score reflects the host environment (moderation quality, monetization model, stranger interaction risk). Individual experience scores reflect the specific content. Consumers of the API should be aware that a "Roblox" platform score is not a score for Roblox experiences as a category — individual experience scores vary widely and should be queried separately.

Live-service games with evolving content. Games with weekly content updates, seasonal events, or ongoing battle passes are scored on their current state at review time plus any persistent structural mechanics. Time-limited events are scored under R1 (FOMO) but the content of those events may not be individually reviewed. Live-service titles are prioritised for periodic rescore when mechanics change materially.

Early access games. Early access titles are scored at the state available at review time. Scores may change substantially on full release. LumiKin does not guarantee rescore coverage for all early access titles on release.

Regional content differences. Some games ship with different content ratings by region (e.g., a game rated M in North America may be rated 16+ in Europe with modified content). LumiKin scores the international/default release where possible. Regional variants are noted where known but are not systematically tracked.

AI-generated scores. Some scores are produced by an automated review pipeline using the LumiKin rubric as a prompt. These are marked review_tier: automated in the API. Expert and community reviews are marked accordingly. Automated scores undergo periodic spot-check auditing but have not been individually verified by a human reviewer.

Changelog